一种利用差集的加权频繁项集挖掘算法

A novel approach for extracting weighted frequent itemsets using diffsets

挖掘加权频繁项集是多种数据挖掘应用中的关键问题,为提高传统加权频繁项集挖掘算法的性能,在研究概念格模型和差集Diffsets理论的基础上,构建一种利用差集的加权频繁项集格结构,该格结构通过差集性质快速计算加权支持度,满足向下封闭特性,更有利于高效生成加权频繁项集.最后给出了相应的FWIL-Diffsets构造算法,该算法仅对数据库进行一次扫描,通过性质定理有效减少项集生成中的计算量.研究结果表明:该方法能显著提高生成加权频繁项集的效率.

The mining weighted frequent itemses（WFI） plays an important role in application of data mining.In order to advance the performance of traditional WFI mining algorithm, a new weighted frequent itemsets lattice model is proposed to make the diffsets theory and lattice model research. A diffset theory for the efficient computation of the weighted support for itemsets is described. The model has the property of downward closure and generate WFI efficiently. Then, the algorithm for the construction of WFI（FWIL-Diffsets） is presented. The algorithms only scan the database once and reducing the amount of calculation.The experiment shows that the proposed method can improve weighted frequent itemsets generation efficiency.